gptkbp:instanceOf
|
gptkb:architecture
|
gptkbp:compatibleWith
|
most AMD processors
most Intel mobile processors
|
gptkbp:containsSpecialSubset
|
gptkb:AVX-512BF16
gptkb:AVX-512BW
gptkb:AVX-512CD
gptkb:AVX-512DQ
gptkb:AVX-512F
gptkb:AVX-512IFMA
gptkb:AVX-512VBMI
gptkb:AVX-512VL
gptkb:AVX-512VNNI
gptkb:AVX-512VP2INTERSECT
|
gptkbp:developedBy
|
gptkb:Intel
|
gptkbp:feature
|
512-bit SIMD operations
new instructions for math, logic, and data manipulation
masking support
|
https://www.w3.org/2000/01/rdf-schema#label
|
AVX-512 (select models)
|
gptkbp:introducedIn
|
2016
|
gptkbp:marketedAs
|
gptkb:Advanced_Vector_Extensions_512
|
gptkbp:maskRegisterCount
|
8 K registers
|
gptkbp:partOf
|
gptkb:x86_instruction_set_architecture
|
gptkbp:potentialDrawback
|
increased power consumption
not universally supported across all CPUs
reduced base clock speed when enabled
|
gptkbp:registerCount
|
32 ZMM registers
|
gptkbp:registerWidth
|
512 bits
|
gptkbp:replacedBy
|
AVX2
|
gptkbp:supportedBy
|
gptkb:Intel_Skylake-X
gptkb:Intel_Xeon_Phi_Knights_Landing
select Intel Xeon processors
select Intel Core i7 processors
select Intel Core i9 processors
|
gptkbp:usedIn
|
gptkb:machine_learning
high-performance computing
scientific computing
data analytics
|
gptkbp:bfsParent
|
gptkb:AMD_Ryzen_processor
|
gptkbp:bfsLayer
|
7
|